Conventional treatment of late-life depression (LLD) often requires long trials of several antidepressants before an effective regimen can be found for an individual. This can take many months and is associated with persistent depressive symptoms, an increased risk of suicide, patients dropping out of care, and worsening of medical co-morbidities. This long response time in LLD is one of the most challenging clinical features of LLD. Thus, in the elderly it is particularly important to shorten this window, and to identify as early as possible what medication regimen will be the most effective for an individual. With the recent development of Pharmacologic Functional MRI (phMRI), it has become possible to track cerebral blood flow patterns (as proxies of regional brain activity) in response to a medication. The primary aims of the proposed LLD phMRI study are 1). To characterize the trajectories of the phMRI brain responses over the course of initiation and titration of an antidepressant, and 2). To use these trajectories to identify the earliest time-point when functional brain changes predict treatment response. As an exploratory aim, the phMRI trajectories will be related to other prominent markers of treatment response in LLD, cognition and structural brain changes. The proposed study is a continuation of R01MH076079, now in its 5th year, which has successfully achieved its aims of characterizing the functional neuroanatomy of LLD, using a model of altered connectivity. The hypotheses for this continuation follow from an observation in the current study that the 12-week change in functional MRI (fMRI) activity was more predictive of treatment response, than was baseline fMRI activity. The proposed study follows-up on this observation by measuring fMRI activity at 5 different time-points over the course of treatment to identify the earliest point at which brain activity in LLD can be used to predict treatment response. The fMRI paradigm includes affective reactivity and affective regulation tasks designed to highlight the ventral limbic affective processing circuits (including the rostral cingulate, anterior insula, and amygdala), which are implicated in LLD treatment. Novel image analysis methods, including resting-state calibration of task-related fMRI, will be used to optimize scan-rescan reliability. One hundred LLD subjects will participate. The proposed study leverages the resources of a recently funded large NIH clinical treatment trial of Venlafaxine XR to treat geriatric depression (IRLGREY, PI: Reynolds, R01MH083648). The imaging component of this study is integrated with the ongoing IRLGREY study, and thus initial subject recruitment, characterization, and treatment will be covered by IRLGREY. This study is designed to impact both ends of the translational research path, from characterizing the pharmacologic effects in the brain to the potential clinical use of MRI for personalized treatment of LLD. As far as we are aware, this will be the first study to systematically characterize the trajectory of fMRI limbic activity to an SRI. The study focuses on the elderly because the importance of predicting treatment response in that group has added urgency due to their prolonged response to treatment and high suicide risk.